AD-A259 775 :UMENTATION PAGE J or Apro

1 041118 ft 12/31/92TITLEW AND fUBITL 5. to FUNDING NUMBE StM a"

Lia~~- Tescione an GeresSlfr

1. ECY UE ONY Lave , REORT ATE3. RPORTTYPREPDO RTE NUMBERE

7roy N/1/9 York9 12180-3590

.TITL APNDS UBTGMCITE P4 51EC AES N ORS~)¶. SFONSOING N MBNITR IN

AGNYREPORT NUMBER

Off ice of Naval Reearch800 North Quincy St.Arlington, VA 22217-5000

i11. SUPPLEMENTARY NOTE$

I s S-5T UT'ON AVAILA931L Tý )TA'E%1E'4T *12b. DISTRIBUTION CODE

unlimited we I£VIIII1

E. coli, genetically engineered with a mercury(tI)-sensitive promnoter and the tax leow trm Vibrio flhcheri, wen used asmicrobial sensors for the detection of mercury. Evaluation of this genetic construction was carried out by desectminsa the

effects of various parameters on cell suspensions maintained at constant conditions in a smal vessel. Tbe strongest light

intensities and quickest induction times occured with cells in the mid-oxponential growth phase mainaianed at 250C.

concentrated to W~t09 cells/fmL, mixed at very fast speeds, and aeratesd at 2 vvm (volvo. of aix per volume of culture per

minute) during light ameaurement in the small vessel. The sensitivity of thea" cells to dhe mwuemic iomU lindI the rule of0.02-4 piM (4-900 ppb) and the total response tinte was on the order of one hour, dependingl on the above pesnmalmr. The

cells exhibited great specificity for miercury. The cells have almost equal speciftoity for Organic and isorqnde fornix of the

mnrcuric ion and responded morm weakly to the marcurous ien. A simple. inexpasive. durable misiatmr probe was

constructed and operated using the optimum parameters found in the small vessel as a guide. The mapg of nmajltivlty 10 the

mercuric ion detected in the probe was 0.01-4 I&h4 when aeration was provided.

'I iijBdCT TERNAS IS. NUMBER OF PAGES

metal ion biosensor, bioluminescence, and mer-lux genetic 16. PRICE COOIIconstruction, and mercuryI i ECoRiTV CLASSIFCATION is SECUR!Tv CLASSIFICATION Is. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT

OF REPORT OF THIS PAGE OF ABSTRACT

unclassified Iunclassified Iunclasstified Unclassified* Standard ;orm 298 (Rev 2-89)

a,.,bed M4n 'd~O tji* S

*1 44,.t pmma

.9. ; 'VJO

v~~r3 v~

s -rt,.~jŽ ,~t

ri `f~ -A . 711

1824

Best Avilbl Copy,

a a a

Tr~~~,A yN A-8

Construction andEvaluation of a Metal Ion

Biosensorgrant # N00014-90-J-1715

Lia Tescione and Georges BelfortBioseparations Research CenterHoward P. Isermann Department

of Chemical EngineeringRensselaer Polytechnic Institute

Troy, New York 12180-3590

Januaiy 1993

93-01804

llll qlllllll qillllq lLIl

Metal Ion Navy Rep.

Construction and Evaluation of aMetal Ion Biosensor

grant # N00014.90-J-1715

by

Lia Tescione and Georges BelfortBioseparations Research Center

Howard P. Isermann Department of Chemical EngineeringRensselaer Polytechnic Institute

Troy, New York 12180-3590

tfor

Dr. Michael MarenONR Code 1141 A006UtInO for

Molecular Biology Program Ts SRA&Office of Naval Research u,,v-,,,,,.',.1,•d 0

800 North Quincy Street J, !.f. t onArlington, VA 22217-5000

"Awvl•.mbilitty Cod#$.._

January 1993 !A Co n.dfa--J IDC i Spec<r I Dl.

D"'r QA rry" In"1"1'11

1. OBJECTIVE.

The objective of this study was to evaluate various growth

conditions on the light emitted from luminescent Escherichia coli and

to construct instrumentation to maintain these cells viable in a

practical, rugged sensor design for the detection of mercury.

2. SUMMARY.

E. coil, genetically engineered with a mercury(Il)-sensitive

promoter and the lux genes from Vibrio fischeri, were used as

microbial sensors for the detection of mercury. Evaluation of this

genetic construction was carried out by determining the effects of

various parameters on cell suspensions maintained at constant

conditions in a small vessel. The strongest light intensities and

quickest induction times occured with cells in the mid-exponential

growth phase maintained at 280C, concentrated to lxl09 cells/mL,

mixed at very fast speeds, and aerated at 2 vvm (volume of air per

volume of culture per minute) during light measurement in the small

vessel. The sensitivity of these cells to the mercuric ion lied in the

range of 0.02-4 gM (4-800 ppb) and the total response time was on

the order of one hour, depending on the above parameters. The cells

exhibited great specificity for mercury. The cells have almost equal

specificity for organic and inorganic forms of the mercuric ion and

responded more weakly to the mercurous ion. A simple,

inexpensive, durable miniature probe was constructed and operated

using the optimum parameters found in the small vessel as a guide.

The range of sensitivity to the mercuric ion detected in the probe

was 0.01-4 gM when aeration was provided.

3. APPROACH.

We have focused on parameter variation in a minireactor test

cell and a design for a miniature probe construction. The minireactor

was a small 100-mL glass vessel that provided mixing, temperature

control, and aeration to a suspended culture, and that was adapted

with a fiber optic to pick up and transmit the emitted light to a

detector.

Conditions of the cell suspensions could be controlled and

parameters varied independently to study their effect on cell light

emission. Cell suspensions were grown in culture flasks, spun down,

and then resuspended and studied in the minireactor. This approach

was easier than using the dual reactor test cell previously proposed

that requires operation of a fermentor. The minireactor incorporated

significant flexibility such that the parameters originally proposed

for investigation in both the Rose chamber test cell and the dual

reactor test cell can be studied in the minireactor alone. Optimum

light intensities and induction times were found for each parameter

that was varied. These optima were used as a guide in operating a

3-mL miniature probe.

4. EXPERIMENTAL PROCEDURES.

4.1 Microbial Strain

The clones were supplied by David Holmes (Clarkson University,

Potsdam, NY). The mer operon from Serratia marescens and the lux

2

genes from Vibrio fischeri were fused and expressed in Escherichia

coli (strain JM109).

4.2 Cell Cultivation

For all studies, Escherichia coli containing the mer-lux plasmid

were grown in a Luria broth medium (LB) with 50 gg/mL ampicillin

(Sigma, St. Louis, MO). The LB medium was prepared by adding the

following (in grams per liter of distilled water): bactotryptone (Difco,

Detroit, MI), yeast extract (Sigma), sodium chloride (Mallinckrodt,

Paris, KY), and glucose (Fisher, Fairlawn, NJ). Precultures were

inoculated with a toothpick by picking a colony grown on an agar

plate of the same medium. Precultures were grown overnight with

shaking in 12 mL of medium. These precultures were used to

inoculate culture flasks. Cultures were grown with shaking in 88 mL

of medium. Except where noted, cultures were inoculated with 20%

(v/v) inoculum. All culture transfers were carried out in a UV-

sterilized culture enclosure (Labconco, Kansas City, MO). Cell counts

were done by serial dilution and absorbance measurement at 660

nm.

Mercuric chloride (Fisher), except when varied, was used at a

concentration of 0.2 pM. The other analytes were mercurous

chloride, zinc(II) chloride, ferrous chloride, cupric chloride, cobaltous

chloride, aluminum(III) chloride, and mercuric acetate (Aldrich,

Milwaukee, WI).

3

4.3 Cell luminescence characterization studies

Sample Preparation. Cultures were grown for 3 1/2 hours,

promptly removed, and separated into 25-mL portions for

centrifugation. The cells were centrifuged at 160 C for 15 minutes at

3000 rpm. After centrifugation, one portion was used per

luminescence measurement. The portions were used as needed for

measurements over a 12-hour period, unless otherwise indicated.

The centrifuged cells remained at the room temperature of 220C.

Only four measurements could be made from one 100-mL

culture. During the course of an experiment, however, more than

four measurements were always taken, so more than one culture was

needed. Generally two cultures and four precultures were used. In

all the experiments, the cultures were inoculated with equivalent

volumes from the four precultures and all centrifuged portions

contained equivalent amounts from both cultures. This method of

volume allocation or proportioning was adopted to minimize

variations from measurement to measurement resulting from

possible variations from culture to culture.

Light was measured in a 100-mL jacketed glass vessel (Kontes.

Vineland, N.J.), the "minireactor." The centrifuged cells were

resuspended and added to the minireactor. Cell counts were taken

for each sample.

Light detection equipment. The minireactor was adapted

with inlets and outlets for addition and removal of culture, analyte,

and sterile rinse water. A glass tube was fitted into the central port

4

of the minireactor for insertion of a fiber optic (S:hott Fiber Optics

Inc., Southbridge, MA) (core diameter = 64 mm, length = 60 cm, NA =

0.56, glass core). The effective volume of the acceptance cone that

the fiber optic saw was approximately 8 mL. The minireactor was

covered with an opaque cloth to remove ambient light interference.

The temperature was controlled with a water jacket and maintained

at 280C, except where noted otherwise. The temperature was

measured with an RTD (Omega, Stamford, CT) and recorded using a

personal computer. A magnetic stir bar provided mixing. In some

experiments aeration was provided through a sparger. The air was

sterilized with a 0.2A filter (Gelman Sciences, Ann Arbor, MI). The

flowrate was monitored with a rotameter (Cole-Parmer, Chicago, IL).

A schematic is shown in Fig. 1.

The fiber optic transmitted the luminescence to a R1527

photomultiplier tube (Hamanmatsu, Bridgewater, N.J.) powered by a

high voltage power supply (Pacific Instruments, Concord, CA) set at

960 volts. The electrical current from the PMT was sent to a 2A50

amplifier (Pacific Instruments, Concord, CA) with an output

impedance of 50 Q and a gain of 100, and emanated with an

associated voltage. The voltage was recorded with a data-acquisition

software (Unkelscope, Cambridge, MA) for one hour via a DAS-8PGA

A/D board (Keithley-Metrabyte, Taunton, MA). This particular set-

up gave a current (mV) to light (photons/sec) conversion of 1.25x10 5

photons/mV.sec. The efficiencies and conversion factors for the

different pieces of equipment are given in Table I.

5

4.4 Probe miniaturization studies

Cells were grown and prepared as described above but

centrifuged in three-mL quantities. Approximately 2.5 mL of a

resuspended sample were added to a Union Cross fitting (Swagelok,

Solon, OH) adapted as shown in Fig. 2 and the luminescence

measured as described above.

A glass disk was glued into the right-hand side fitting of the

cross, the "probe." The fiber optic was fitted with ferrules into this

fitting. The glass disk protected it from contact with the culture. The

effective volume that the fiber optic saw in the probe was

approximately 1.4 mL. The whole probe was kept in a 280 C water

bath. A miniature stir bar provided mixing. Aeration was provided

through the left-hand side fitting; the air was filtered and its rate

measured as done in the minireactor. The top fitting was cracked

open for air to escape, while the bottom one was scaled tight with a

Plug nut (Swagelok). The probe was light-proof.

S. RESULTS AND DISCUSSION.

5.1 Approach

Several parameters were varied in the minireactor: culture

age, temperature, cell density, mercury concentration, mixing rate,

6

Table I. Efficiencies and conversion factors for components of the light-

detection set-up.

Instrument Parameter Value

Fiber Optic effective transmission (%), T 40

Photomultiplier Tube (PMT) quantum efficiency (%), Qa 10current amplification, A 4W10 5

A/D Board 50impedance (Q), R 100gain, G

a) Q = number of photons incident onto the PMT/ number ofphotoelectrons emitted from the PMT. The current (1) is calculated asfollows

P (photos/sec) * 1.6x10 1 9 (Coulomb/electron) * A '0 * R (0) 11 1000 in V/VI (iV) "- Q (photons/electrons) * T

- x10"6 (mV/photons/sec) * P (photons/sac),

where the light-to-current conversion is 8x10"6 mV sec photon-landits inverse, the current-to-light conversion, is 1.2 5x105 photon sec-ImV- 1.

7

and aeration rate. Optimum values were found for the strength of

the light emission (light intensity) and the induction time. These

optima were used as a guide for operating the minisensor. A

correlation with the concentration of mercury was obtained in the

minisensor.

5.2 Results

5.2.1 Signal characteristics.

The cells were found to emit light approximately thirty

minutes after addition of mercury. This time lag, called the induction

time (I), indicates when the emitted light was first detected. The

cells did not emit a flash of light at the induction time and then cease

luminescing. Rather, the emission continued to increase for at least

another 30 minutes.

Since the signal kept increasing, one had to determine when

the cells will gave a sensitive correlation to the concentration of

mercury. This time is called the response time (r). The total

response time (R) of these microbial sensors is the summation of the

induction time and response time. The light emission was not

recorded for more than one hour in the minireactor because total

response times of more than one hour were discarded as too long.

The signal strength in the minisensor was weaker than that in

the minireactor because of the smaller volumes used, so light

emission was monitored for 1 1/2 hours in order to obtain stronger

light intensities. Longer total response times were required in the

minisensor to obtain a sensitive correlation.

8

5.2.2 Parameter variation In the minireactor test cell.

Before commencing optimization studies of the operating

parameters used in the minireactor, the phase at which the cells give

off maximum light needed to be found. The cells were grown for

their entire growth cycle trom the lag phase to the stationary phase.

At different intervals 25 mL of the culture was removed, added to

the minireactor, and the luminescence measured as previously

described. The culture was not centrifuged prior to measLrement.

The growth curve of the cells for two different percent inocula is

shown in Fig. 3 and the light emission during the course of the

growth cycle is shown in Figs. 4a and 4b for 1% and 20% (v/v)

inoculum, respectively. The cell density (Figs. 4c and 4d) and the

calculated light output per cell (Figs. 4e and 40 are also shown for 1%

and 20% (v/v) inocalum. The point of maximum light emission

occurs with the 20% (vWv) inoculum culture at 3 1/2 hours into the

growth cycle, so for all future studies cells were taken from a 20%

(v/v) inoculum culture that had grown for that length of time.

High cellular luciferase contents must exist during the early

growth phases because of the high light output per cell (Figs. 4e and

4f). For the 1% (v/v) inoculum culture, a second peak occurs during

stationary phase (Fig. 4b) because of the higher cell densities.

The effects of temperature on light intensity and response time

are shown in Fig. 5. Maximum light intensity occurs at 260C and

minimum induction time occurs at 300C, so 28 0 C was selected as a

"trade-off temperature for all future studies.

9

The effects of cell density on light intensity and induction time

are shown in Fig. 6. The light intensity increases with more cells

present, but above a certain cell density, it drops off probably due to

the scattering of the light by the cells.

A correlation of concentration of mercuric ion with light

emission is shown in Fig. 7. The most sensitive correlation was

obtained with a 30 minute response time. Longer response times

were not considered, as mentioned above. These cells exhibited

sensitivity for the mercuric ion in the range 4 ppb to 200 ppb (0.02-

4 jtM). The toxicity of mercury began to adversely effect the light

e, nission from the cells above 20 ppb because of the absence of a

complete merA gene. Light emission was nearly linear with

concentration for one order of magnitude from 0.04 to 0.4 gM (20

ppb), where it peaked. For higher concentrations light emission was

nearly linear, but inversely proportional to concentration.

Bioluminescent E. coil which have not been cloned with a toxin-

specific promoter exhibited bioluminescence quenching in the same

range, 0,4-2.0 gtM (1). Metabolic inhibition in E. coil started at a

critical mercury concentration of 0.4 ýtM.

The mixing rate was not found to affect light intensity,

although it did influence induction time. Induction time decreased

linearly with mixing rate (Fig. 8). This trend indicates that the rate

of diffusion of the mercury into the cells affected the response time

of the cells and that faster stirring speeds could be used to shorten

total response time of the microbial sensors.

Aeration increased the light intensity by 50% (Fig. 9). Thus, the

luciferase-catalyzed reaction must he inherently oxygen-limited.

10

Above about an air flow rate of 50 ml/min, the light intensity

decreased possibly due to multiple scattering. Perhaps, E. coil cloned

with bioluminescent genes should also be transfected with genes that

enhance oxygen uptake, such as the Vitreoscella gene constructed by

Bailey et al. (2), in order for sufficient oxygen to be present for the

luminescent reaction. Aeration also caused an increase in induction

time, possibly due to a dilution effect from the additional air bubbles.

Sensitivity to other metal ions was tested. No light was

detected from 0.2 p.M of zinc(ll) chloride, ferrous chloride, cupric

chloride, cobaltous chloride, and aluminum(Ill) chloride. The

mercuric ion was detected at similar light intensities and induction

times for the inorganic and organic (mercuric acetate) forms. The

cells were also sensitive to the less prevalent mercurous ion (Fig. 10);

however, the detected signal was much weaker in both intensity and

induction time.

5.2.3 Cell Pellet Age.

Thc long-term storage of the stock cultures used for streaking

plates may have affected the luminescent behavior of cultures

inoculated from these stock cultures. The cells initially gave

reproducible signals for the duration of the experiments. However,

after long-term storage (more than 60-75 days) of the stock cultures

at -20oC and several thawing/refreezing cycles, changes in light

emission were observed. The light intensity and induction time of

these cells did not remain constant while they remained at ambient

temperatures as a pellet (after centrifugation) (Figs. Ila and lib), as

they did when younger stock cultures were used. Within six hours,

11

LU

induction time droped precipitously and then leveled off. Light

intensity increased with cell pellet age, the time the centrifuged cells

remained at ambient temperatures, and levelled off after 8-9 hours.

An interim ten-hour period, when the pellets are 10-20 hours old,

existed in which light intensity and induction time did not vary

significantly (Fig. 11).

The cells used for all the studies described in the previous

section (except aeration rate and other metal ions) were grown from

stock cultures that were less than two months old. The light

intensities and induction times for those studies did not follow the

trends in Figs. Ila and llb with cell pellet age, so they are not

invalidated by this effect. However, later experiments, including

repetition of the variation of the aeration rate in the minireactor, test

of the sensitivity to other metals, and all studies in the minisensor

(see below), were carried out using cell pellets that were 10 to 20

hours old in order to minimize variation resulting from storing the

pellets at ambient temperatures.

The effect of mixing rate was repeated at a later date using

cell pellets that were 10-20 hours old, and it was found that mixing

rate decreased linearly at neardy the same rate (slope) as before.

This also indicates that the previous trends were not artifacts of cell

pellet age. However, the curve shifted downward; the induction

times were quicker when the stock culture used had been stored for

a longer period of time (60 days longer) and undergone more

freezing/thawing cycles (Fig. 8). Since cell luminescence behavior

differs with storage time, the cells should be tested periodically and

the probe calibrated before use.

12

I II II • . . . -

5.2.4 Minisensor evaluation.

Measurements in the minisensor were taken using cells with a

pellet age of 10-20 hours. The mixing rate was approximately 3200

rpm. No light was detected in the minisensor without aeration (Fig.

12). An aeration rate of 15 mL/min or 6 vvm (volume of air per

volume of culture per minute) was used because it gave the

strongest light intensities in the minisensor. This value differs from

the optimum rate found in the minireactor, 50 mL/min (2.0 vvm).

The difference may be due to the fact that a sparger was used in the

air inlet to the minireactor, whereas the air gently bubbled into the

minisensor.

The correlation obtained with concentration of mercury in the

minisensor is shown in Fig. 13. The detectable range of sensitivity to

the mercuric ion is almost the same as that obtained in the

minireactor, 0.01-4 g±M, and parabolic in nature; although the

sensitivity at the low end of the range behaved differently. The

induction time was on the order of 30 minutes, as it was in the

minireactor. However, a more sensitive correlation with

concentration was obtained with a response time of 60 minutes than

that of 30 minutes, but the total response time then was one and

one-half hours instead of one hour. The peak shifted to 1.0 gM with

longer response times. It must be realized, when comparing the

correlations sho,.-n in Fig. 7 and Fig. 13 for the minireactor and

minisensor, respectively, that there was no aeration for the

correlation obtained with the minireactor, the mixing rate was much

13

less in the minireactor, and the effective volumes of culture that the

fiber optic sees in the minireactor and minisensor are different.

6. CONCLUSIONS.

Cell density, temperature, mixing, and aeration are all

important factors in maximizing light output of genetically-

engineered bioluminescent Escherichia coll. Cells in the mid-

exponential growth phase produced the strongest light intensities.

Cultures in this growth phase exhibited optimum trade-off between

light intensities and induction times in the minireactor when

maintained at 280C, concentrated to Wx10 9 cells/mL, mixed at very

fast speeds and aerated at 2 vvm (volume of air per volume of

culture per minute).

The range of sensitivity for the mercuric ion was 0.02-4 gM in

the minireactor and 0.01-4 pM in the minisensor, when aerated. The

correlation of light intensity and concentration of the mercuric ion in

the minireactor was parabolic in nature with regions in the mid-

range exhibiting linearity. The correlation obtained in the

minisensor was similar to the one obtained in the minireactor.

The induction time of the cells was approximately 30 minutes.

After the induction period the light intensity increases. Sensitive

correlations were obtained 30 minutes beyond the induction period,

resulting in a total response time of one hour.

The cells exhibited exquisite specificity for mercury, in

particular for the mercuric ion. No response was detected from the

other metals ions tested. Since all forms of mercury are toxic, these

14

cells may prove useful in toxicity analyses and certainly require

simpler protocols than those of the cold vapor methodology that are

used when different forms are present.

Light emission was studied from cell suspensions. It was found

that luminescence may vary with short-term durations (on the order

of hours, the duration of an experiment) after long-term storage (on

the order of months) of the stock cultures at -20oC. When using

these cells in a probe for field testing it will be important to

periodically check the cells' luminescence and to calibrate the probe

prior to use. Further investigations should be carried out to test how

luminescence changes over long periods. Storage at temperatures

lower than -20oC may reduce variability in the light emission from

the cells.

Studies need to be carried out to investigate how to maintain

the cells in a static state. Hank's balanced salt solution (HBSS) has

been found to stabilize the light signal from cell suspensions (1).

HBSS may also enhance sensitivity because the solution is clear and

less !ight will be scattered.

Other phases, such as lyophilization or immobilization, should

be investigated for maintaining the cells in a static state. If cell

suspensions cannot be maintained in a static state for a substantial

period of time and will requirc on-site culturing, then these other

phases may make eventual operation of a probe easier than cell

suspensions. Variation of luninescence over time shotild be tested

for these phases as well as the effects of storage temperatures

(refrigerator, freezer, etc.) on luminescence. Akerman et al. (3)

observed changes over time in immobilized bioluminescent E. coli

15

and they also noticed a difference with the method of storage of the

beads.

7. RECOMMENDATIONS.

The trends observed for the different parameters studied can

be used to improve biosensor sensitivity and response time and they

can serve as a guide for probe design. Since temperature is a

variable that influenced light emission, temperature control is

important, but it may be difficult to obtain in the field. Providing for

thermostasis with a water bath, even if small, is inconvenient.

The sensitivity of the cells and the correlation with the

concentration of mercury may be further enhanced by concentrating

the cells during resuspension to the optimum cell density of lx109

cells/mL. Since the correlation was parabolic, dilution may be

necessary during operation in order to know which half of the

correlation curve is being detected. Although, if the entire mercury-

detoxifying merA gene could be inserted, the toxicity effects and

parabolic correlation may be eliminated and the linear region

extended into the range of higher concentrations.

Addition of exogenous oxygen is important because the cells

are oxygen-limited. Incorporation of the Vitreoscella gene may

enhance oxygen uptake and cell response. The effect of the addition

of oxygen-carrying molecules such as hemoglobin or

perfluorocarbons may also be evaluated; although, this approach may

be less convenient because additional components need to be added

whereas the gene would be an inherent part of the biosensor.

16

In the field, if the cells are found to be oxygen-limited even

with the Vitreoscella gene and/or oxygen-carrying molecules, air

may be easily added to the probe with an inexpensive aeration

pump. The effects of ambient air teioperature, pressure, and

humidity would have to be investigated. The Vitreoscella gene may

also be useful if immobilized cells are used because there are even

greater diffusion limitations in gel beads than in free suspension.

Since oxygen is a limiting component, limitation of other

substrates may need to be tested. Addition of the other substrates,

i.e., aldehyde and FMNH2, may also improve cell response.

Other parameters that should be tested for their effect on cell

light emission include pH, other culture media (i.e., M9 minimal

medium), and the use of transparent solutions for resuspension of

the centrifuged cells (minimal medium, salt solutions such as HBSS,

buffers).

Greater mixing shortened cell induction time and improved

diffusion of mercury into the cells. In the minireactor the sparger

broke up air bubbles through convective flow producing more but

smaller bubbles for increased surface area and diffusivity. It also

provided additional mixing and turbulence. The minisensor,

however, had no sparger. If a membrane, filter, or mesh could be

inserted at the inlet of the probe, air could be forced to diffuse into

the probe inducing sufficient mixing such that a stir bar may not be

necessary in the minisensor. This would eliminate the need for a

magnetic stirrer, increases the volume for the cells and further

simplify probe operation.

17

The probe design may be adapted for flow.through if

continuous measurements need to be made. The cells and samples

tested could be compartmentalized with membranes. Although, an

alternate design may be warranted.

Microbes are able to detect high levels of mercury

contamination in water. The genetic construction used in this work is

more sensitive than other microbial detectors. It is slightly less

sensitive than cold vaporization atom'ic absorption spectrometry,

based on values reported by the EPA, and it cannot yet qqtmpete with

some newer commercially-available detectors whici'*have even

lower sensitivities (although, they do not indicate the working linear

range) than those reported by the EPA.

The sensitivity of these toxin-specific microbes may be further

lowered with some of the aforementioned suggestions (elimination of

substrate-limitation, suspension in HBSS, concentration to higher cell

density, etc.) and perhaps the parabolic correlation eliminated by

altering the genetic design.

In conclusion, the studies carried out in this investigation, those

proposed for future work, and the probe design presented may serve

as a prototype for studying othcr bioluminescent cells and

developing biosensors using such cells. Thus, a family of

bioluminescent metal ion biosensors could be built based on the

ideas presented here for the genetic design and methodology for

construction and evaluation of a probe.

18

8. APPENDIX.

A sample measurement of a light signal recorded with the

light-detection equipment described earlier is shown below (Fig. 14).

Light intensities (in millivolts) and induction tirn-s were extracted

from such signals. The induction time (I) is indicated. As previously

defined, I is the duration for which no bioluminescence is detected

and only the background level is measured. The signal coming from

the photomultiplier tube is negative, so a negative deviation from the

background represents light emitted from the cells. A line was

drawn through the lowest values that follow the general trend of the

signal; scatter data points were disregarded. Values for light

intensities were calculated as the difference of the background value

and the values falling on the drawn line for a selected response time.

9. PUBLICATIONS AND PATENTS.

A manuscript resulting from this research has recently been

submitted for publication:

Tescione, L. and Belfort, G., "Construction and Evaluation of a

Metal Ion Biosensor" submitted to Biotechnology and

Bioengineering, January 1993.

No patents were filed as a result of this research.

19

ACKNOWLEDGEMENTS.

Dr. David Homes supplied the cells. Dr. Holmes, Dr. Marcy

Osgood, and Dr. Santosh Dubey provided consultation. Dr. Fritjof Linz

and Dr. Joel Plawsky assisted in the initial set-up of the light

detection equipment. Hillary Bollam and Mark McGann provided

their assistance in the laboratory.

REFERENCES.

1. Lamplnen, J., Korpela, M., Saviranta, P., Kroneld, R.. Karp, M.1990. Toxicity Assessment, 5: 337.

2. Bailey, J.E. 1991. Science, 252: 1668.

3. Akerman, K.E., Kukkonen, J., and Karp, M. 1990. pp. 173-180 In:J.A.M. de Bont, J. Visser, B. Mattiasson, and J. Tramper (eds.),Physiology of Immobilized Cells, Elsevier Science, Amsterdam.

20

LIST OF FIGURES

Figure I. Small vessel ("minireactor") in which bioluminescence was measured.

Figure 2. Miniaturized probe: Union Cross fitting (Swagelok, Solon, OH) adapted foraeration, mixing, and light detection. Temperature was kept constant at 280 Cwith a water bath.

Figure 3. Growth curves for the cells when grown in a 1% and 20% (v/v) inoculumculture.

Figure 4a) Light intensity from a 1% (v/v) inoculum culture and b) a 20% (v/v) inooculumculture from two experiments; c) cell density from the 1% culture and d) 20%culture from two experiments; and e) the light output per cell for the 1% cultureand f) 20% culture from the two experiments.

Figure 5. Effect of temperature on light intensity and induction time of the cells.

Figure 6. Effect of cell density on the light intensity and Induction time of the cells.

Figure 7. Correlation of light intensity with different concentrations of mercury fordifferent total response times. Dashed lines suggest a possible linearrelationship.

Figure 8. Effect of mixing rate on induction time using cells with a pellet age of 0-12hours old (.) and using pellets 10-20 hours old (o).

Figure 9. Effect of aeration on light intensity and induction time,

Figure 10, Light intensities measured after induction with various metals.

Figure 11. Effect of pellet age on cell bioluminescence: a) light intensity, b) induction time,and c) cell density.

Figure 12, Effect of aeration on light intensity in the minisensor (response time, r, is 60minutes).

Figure 13, Correlation of light intensity in the minisensor with different concentrations ofmercury for different total response times.

Figure 14. Sample raw data measurement and extraction of light intensity and inductiontime (I).

TT Fiber Optic

Culture, Giles Portanalyte

Drain

Air "

Wator

Figure 1. Small vessel ("rminireactor") in which biolumiilescence was

measured.

Stirring GlassBar Ois

Fiber Optic

Air

Water Bath

I0 0 I Magnetic Stirrer

Figure 2. Miniature probe: Union Cross fitting (Swagelok, Solon, OH)adapted for aeration, mix-1g, and light detection. Temperature was

kept constant a, 28 0C with a water bath.

2

2 10.

ZI

108

0 1 2 3 4 5 6 7 8 9 10 11 12

Cultivation Time (hr)

Figure 3. Growth curves for the cells when grown in a 1% and 2,0%(v/v) inoculu0 culture.

21 --- -i a) 23 0 720. 2C

ate .

i_ T

10.4

109

atol

2 3 T "'Cw towrr .'"

_ , a I .. ... ,

I • I 110-4

L

1011

z a 2

Figure 4a) Light intensity from a 1% (v/v) inoculum culture and b) a20% (v/v) inooculum culture from two experiments; c) cell densityfrom the 1% culture and d *) 20% culture from two experiments; and e)the light output per cell for the 1% culture and 0 20% culture fi omthe two experiments.

5540

3

E

. 20 35

0' I10 025

.0

,i •

24 26 28 30 32 34 36 38Temperature (C)

Figure 5. Effect of temperature on light intensity and induction timeof the cells.

30 50

d=

* 0 45.4- .' e-

E 20 -

4 40A E.lo

3010 1 .. "

0... •" 30

0 107 2 3 4 10J 2 3 4 lo 2 3 4 101

Cell Density (cellu/mL)

Figure 6. Effect of cell density on the light intensity and inductiontime of the cells.

50 1**=* + 10rnim

~20

10

?oY-2 2 354S 10-1 2 3 45 1(00 2 3 4

Figure 7. Correlation of light intensity with different concentrationsof mercury for different total response times. Dashed lines suggest apossible linear relationship.

41

37

5 35

3.3S31

.2 0*• 29

27

250 .

23 021,

0 1000 2000 3000 4000Mixing Rate (rpm)

Figure 8. Effect of mixing rate on induction time using cells with a

pellet age of 0-12 hours old (s) and using pellets 10-20 hours old (o).

F1

SIU

110 31

30

"> ... .. .........

2990

28.

27

70 60 20 4.0 60 80 100 120 14P

Air Flowrote (mL/m in)

Figure 9. Effect of aeration on light intensity and induction time,

60

50

E 40

' 30

1 20

10

. ... I I I

Al(Ill) Co(lO) Cu(') Fe(Il) HgOAc Hg(II) Hg(I) ZnM(l)

Metal ion

Figure tO. Light intensities measured after induction with varouS

metals.

£ •. .

70

160

30

10 3

29

27

0 5 13 15 2'0 '25 30r'"16e SOmol "Ofmaine at armbiermt temperature (hr)

109

lost10 5 20 0 5 20 as 30

r:-o sarmos romagimed at amollent temp (hr)

Figure 11. Eff ct of pellet age on cell bioluminescence: a) lightintensity, b) induction time, and c) cell density.

70

60 / 90 .... "'

50

.~301

20

10

0 10 20 30 40 50 60 70 80 90 100 110 120

Air Flowrote (mL/min)

Figure 12. Effect of aeration on light intensity in the minisensor

(response time, r, is 60 minutes).

601-o- +1 0 min / ,..

50 .... i,. 1+20 min

.-,,- 1+30m i / PiI.in,.,, .0--..- 140ma /m/S40 1+50 min.*... .- 1+60 mrin \ \

.3030

- 201P /, .*.,.. .,.*

10 __ _ _ _ _ __ _ _ _ _ _ __ _ _ _ _ _ __ _ _ _ _ _

10-2 2 45 101 2 354 100 2 3*5

Figure 13. Correlation of light intensity in the minisensor withdifferent concentrations of mercury for different total responsetimes.

* . - - -I..

I I I

Figure 14. Sample raw data measurement and extraction of lightintensity and induction time (I).

Distribution List for Final Reports

Attach a copy of the REPORT DOCUMENTATION PAGE (DD FORM 1473) to your final reportas the first page and mail two copies (including the postcard labelled DTIC FORM 50 ) to:

Defense Technical Information CenterBuilding 5, Cameron StationAlexandria, VA 22314

This allows other investigators to obtain copies of your report directly from DTIC.DTIC will fill out the postcard DTIC ACCESSION NOTICE (DTIC FORM 50) and return it toyou with their number for your report. When you refer people to DTIC to get a copy ofyour report, give this number to expedite the request.

Mail one copy to each of the following and attach this very page to the back of yourreport - otherwise the folks below will think they have mistakenly received a copymeant for the Molecular Biology Program):

(a) Dr, Michael Marron (e) DirectorONR Code 1141 Chemical and Biological Sci DivMolecular Biology Program Army Research Office800 N. Quincy Street P. 0. Box 12211Arlington, VA 22217-5000 Research Triangle Park, NC 27709

(b) Administrative Grants Officer () Life Sciences DirectorateONR Resident Representative Air Force Office of Scientific Res(address varies - see copy of your Bolling Air Force Basegrant/contract) Washington, DC 20332

(c) Director, (g) DirectorApplied Research Directorate Naval Research LaboratoryONR Code 12 Technical Information Div800 N. Quincy Street Code 2627Arlington, VA 22217-5000 Washington, DC 20375

(d) DirectorOffice of Naval TechnologyCode 22800 N. Quincy StreetArlington, VA 22217-5000

Encl (1) 3/12/91